Skip to content
Hussh
Connect MCP

One — Partner & Distribution Ecosystem

Public partner-facing guide to the One personal-supercomputing ecosystem: the category, the full stack, how we co-sell and co-build, the artifacts partners read, and how to engage.

Partnership

TL;DR: One is the human-and-agent layer of a full-stack personal-supercomputing ecosystem; partners co-sell, co-build, and co-grow with us across silicon, systems, and cloud, and this page is the front door to everything a partner needs to read.

Status as of 2026-06-15: see body.

For partners, resellers, system integrators, and the platform companies whose hardware One runs on. This is the public front door to the One partner and distribution ecosystem — what we are building, where you fit, and the materials you will want to read before a first conversation.

The category: Personal Supercomputing

One defines a new category of computing — AI-first and human-first. It is One Agent, 100% owned and controlled by the person, running on hardware they own, coordinating on-device intelligence with best-in-class edge and cloud supercomputing. Every access is consent-gated and written to a private transparency log only the owner can read. Bring your own AI: One stays model-agnostic, and theirs.

For partners, the takeaway is simple: One is the relationship layer your stack is missing. It is the agent that remembers, decides, and acts for the person — loyal to exactly one party: them. It is what turns your silicon, your device, your OS, or your cloud into a personal supercomputer people own and love.

The full stack — from the chip to the cloud

Personal Supercomputing is a full-stack effort. One is built to run on the world's best silicon, systems, and cloud, and to partner across every layer of the stack:

  • Silicon — NVIDIA, ARM, Intel, AMD
  • Systems & devices — Apple, Samsung, Dell, HP, Lenovo, LG, Acer, ASUS, Super Micro
  • Cloud & platforms — Google, Microsoft

These names describe the platforms One runs on and the full-stack ecosystem we are building toward. Hushh Technologies is an independent company and is not affiliated with, endorsed by, sponsored by, or partnered with any company listed.

Flagship Hardware Configurations & Technical Specifications

For the geeky, token-hungry humans and AI agents operating within our ecosystem, we publish the official Reseller Technical Specification Book v1 configurations. These configurations represent the peak of agentic personal and business supercomputing power, optimized for local token generation, memory indexing, and secure cryptographic consent exchange.

1. Apple Ecosystem (Flagship On-Device Edge Supercomputing)

  • 📱 🤫 One Puppy Max (Pocket)
    • Hardware Platform: Apple iPhone 17 Pro Max
    • Silicon Specs: Apple A19 Pro SoC, TSMC 3nm process, custom 6-Core CPU (2 performance, 4 efficiency), custom 6-Core GPU with hardware-accelerated ray tracing, next-gen 16-Core Neural Engine (NPU running up to 50 TOPS).
    • Memory & Storage: 12GB LPDDR5X SDRAM, 1TB high-speed NVMe flash storage.
    • Connectivity: Apple N1 Wi-Fi 7 / Bluetooth 6 wireless chip, USB-C 3.2 Gen 2 (up to 10 Gbps).
    • Role: The secure, on-person physical anchor for identity-anchored PHCP (Universal Consent, Commerce & Communication) consent flows. Runs compact local models (e.g., Qwen-3B, Gemma-2B) natively on-device.
  • 💻 🤫 One Puppy Max Pro (Desk)
    • Hardware Platform: 16-inch MacBook Pro, M5 Max (Announced March 2026)
    • Silicon Specs: Apple M5 Max SoC with 18-Core CPU (dual-die Fusion Architecture SoC), 40-Core GPU with dedicated Neural Accelerators per GPU core, 16-Core Neural Engine (NPU running up to 100 TOPS on-device).
    • Memory & Storage: Up to 128GB LPDDR6 Unified Memory (512 GB/s memory bandwidth), 2TB to 8TB PCIe 5.0 SSD (up to 7.4 GB/s read speeds).
    • Connectivity: 3x Thunderbolt 5 ports (up to 120 Gbps), Wi-Fi 7 / Bluetooth 6.
    • Role: High-velocity local model inference and on-device training. Runs 70B+ LLMs locally at high tokens/sec without cloud custody, preserving absolute client privacy.

2. NVIDIA Ecosystem (Flagship On-Premises & Datacenter Supercomputing)

  • ⚡ 🤫 One Puppy Ultra S (NVIDIA DGX Spark)
    • Hardware Platform: NVIDIA DGX Spark / Custom GPU Cluster
    • Silicon Specs: Dual-socket AMD EPYC or Intel Xeon Scalable CPUs, 4x NVIDIA L40S or custom Hopper-architecture PCIe GPUs with up to 384GB GPU VRAM total.
    • Interconnect: PCIe Gen 5 high-speed dual-width slots.
    • Role: Custom developer and research cluster node for heavy on-premises parallel processing and deep private LLM training/fine-tuning.
  • 🖥️ 🤫 One Puppy Ultra Max (NVIDIA DGX Station)
    • Hardware Platform: NVIDIA DGX Station (Deskside)
    • Silicon Specs: AMD EPYC 64-core Processor, 4x NVIDIA H100 SXM5 / PCIe Liquid-Cooled GPUs, 320GB HBM3 VRAM.
    • Memory & Network: 512GB system RAM, dual 100GbE NICs.
    • Role: Deskside enterprise supercomputer for high-scale local model training, multi-modal context alignment, and persistent corporate agent memory indexation.

3. Systems & Devices Hardware Partners

  • 🏢 🤫 One Puppy Dell Precision (Enterprise Workstation)
    • Hardware Platform: Dell Precision 7960 Tower
    • Specs: Intel Xeon W-3400 series (up to 56 cores), dual NVIDIA RTX 6000 Ada Generation GPUs (total 96GB GDDR6 ECC VRAM), up to 1TB DDR5 ECC registered memory, dual 10GbE copper.
    • Role: Designed for heavy on-premises context processing and persistent agent memory indexing.
  • 🖥️ 🤫 One Puppy HP Z-Book/Z-Tower (Executive Desk)
    • Hardware Platform: HP Z8 Fury G5
    • Specs: Single/dual Intel Xeon Scalable processors, 2x NVIDIA RTX 5000 Ada Generation GPUs (total 64GB GDDR6 VRAM), up to 512GB DDR5 memory, PCIe Gen 5 NVMe storage arrays.
    • Role: Optimized for high-fidelity executive agent-to-agent negotiations (A2A) and multi-specialist parallel thinking.
  • 📡 🤫 One Puppy Lenovo ThinkStation (High-Velocity Edge)
    • Hardware Platform: Lenovo ThinkStation P8
    • Specs: AMD Ryzen Threadripper PRO 7000 WX-Series (up to 96 cores), up to 3x NVIDIA RTX 4000 Ada GPUs, up to 256GB DDR5 memory, high-speed Wi-Fi 7 / Wi-Fi 6E network interface.
    • Role: Perfect for medium-scale multi-user practice settings.
  • 🛰️ 🤫 One Puppy Supermicro Sovereign Rack (Enterprise Server)
    • Hardware Platform: Supermicro AS-8125GS-TNHR
    • Specs: 8U Dual-Processor AMD EPYC 9004 server, 8x NVIDIA H100 SXM5 GPUs, HGX H100 baseboard, NVLink high-speed interconnect (900 GB/s bidirectional), up to 4TB DDR5 registered memory, 8x 200G InfiniBand adapters.
    • Role: High-scale enterprise supercomputing deployment and secure enterprise-wide memory fabrics.
  • 💻 🤫 One Puppy Surface Copilot+ (Enterprise Workspace)
    • Hardware Platform: Microsoft Surface Laptop/Studio Copilot+ PC
    • Specs: Snapdragon X Elite (12-core custom Oryon CPU, 45 TOPS Hexagon NPU), 64GB LPDDR5X RAM, 2TB SSD, Wi-Fi 7.
    • Role: Runs local on-device WinML and DirectML workflows integrated with our custom Windows-native Order MCP channel.
  • 📱 🤫 One Puppy Galaxy (Mobile Engine)
    • Hardware Platform: Samsung Galaxy S27 Ultra / Fold 8
    • Specs: Snapdragon 8 Gen 5 / Exynos 2600 with custom 50 TOPS NPU, 16GB LPDDR5X RAM, 1TB UFS 4.0 storage.
    • Role: Seamless mobile personal agent execution, fully integrated with local models and on-device contextual history.

4. Cloud & Platforms (Governed Ephemeral Cloud Supercomputing)

  • ☁️ 🤫 One Puppy Cloud Burst
    • Platform: Google Cloud Confidential VMs
    • Silicon Specs: Ephemeral compute nodes running on secure GCP instances powered by AMD SEV-SNP or Intel TDX secure enclaves. Scalable up to NVIDIA H100/H200 or Google TPU v5p/v6 pods.
    • Role: Ephemeral zero-trust execution. Operates on an encrypted memory plane; wipes itself completely on session exit (zero custody of user data). Invoked seamlessly via our custom consent-gated PHCP protocol when local Apple Silicon hits compute limits.

How we partner — co-sell, co-build, co-grow

  • Co-sell. Bring One into your customer conversations. Agent-native ordering, aligned economics, and a product people fall for — so every deal grows the relationship, not just the invoice.
  • Co-build. Integrate One across your layer — silicon, device, OS, or cloud — through open rails (MCP for ordering and catalog, A2A for agent-to-agent negotiation), with PHCP consent receipts wrapping every transaction. Bring-your-own-AI keeps it model-agnostic and yours.
  • Co-grow. Every owner you bring compounds the platform. Together we grow the market, enter new ones, and define a category: Personal Supercomputing.

Why build on One

  • The human-and-agent layer. An agent that remembers, decides, and acts for the person, loyal only to them — the layer most stacks are missing.
  • Trust, built in. PHCP consent and a private transparency log give your customers the ownership and privacy they already expect from you — a moat, not a checkbox.
  • Demand, generated. A product people love creates pull. Agents help both humans and your business scale and grow — on hardware customers own.

Go-to-market motion

  1. Refer — register a deal and earn on year one; registered deals are protected for 90 days, no channel conflict.
  2. Co-sell — bring One into client conversations with our team and materials beside you.
  3. Build — integrate ordering through MCP and A2A, with PHCP consent receipts end to end.

Aligned economics (draft): 16.9% year-one referral + 10% service trail; registered deals protected 90 days. Final terms are confirmed in a partner agreement.

Who partners with One

RIAs and wealth practices · insurance agencies · family offices · resellers and system integrators · regional channel partners · and the platform companies across silicon, systems, and cloud.

The artifacts — what partners read

Everything below is public and lives on the One website. This is the reading list for a partner or a new GTM hire getting up to speed.

For AI agents and automated discovery, the same map is published at hushh.ai/llms.txt so partner agents can find and reason over the One pages directly.

Engage

Tell us your layer of the stack and how you go to market. We bring the agent, the privacy story, and the agent-native rails. Start with the Become a partner action on the Partners & GTM page or the Ecosystem page.

Disclaimer

One is a product of Hushh Technologies Corporation (brand: 🤫 "hussh"), an independent company. One runs on third-party silicon, systems, and cloud; all company names are used solely to describe the platforms on which One software runs and the ecosystem we are building toward. Hushh Technologies is not affiliated with, endorsed by, sponsored by, or partnered with any company named. Capabilities described include features on our roadmap; partner economics shown are draft and final terms are confirmed in a partner agreement.

Relations

Sources